Tardis Docs · AI SDK + Runtime API

Build and inspect LLM inference with a request-first SDK and a telemetry-aware Runtime API.

These docs cover the provider-agnostic AI SDK, canonical lifecycle events, ingestion pipeline, and Conversation/Turn APIs.

Open documentationStart with quickstart

Start here

Quickstart

Run the Runtime API locally, create a Conversation, and inspect Inference Request telemetry.

Architecture

Understand how Runtime API, AI SDK, and the DB ingestion layer work together.

AI SDK

Provider adapter contract, canonical stream lifecycle, and normalization behavior.

SDK Examples

Copy-paste examples for complete(), stream(), telemetry hooks, and custom providers.

Runtime API at a glance

List Conversations

/conversations
GET

Create Conversation

/conversations
POST

Continue Conversation

/conversations/:id/messages
POST

Inspect Inference Request

/inference-requests/:id
GET